Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
World Community Grid Forums
Category: Completed Research Forum: The Clean Energy Project - Phase 2 Forum Thread: Poor crunching on this project |
No member browsing this thread |
Thread Status: Active Total posts in this thread: 156
|
Author |
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I also have started seeing problems with CEP2 reported CPU time. I Have Ubuntu 32-bit 10.04, BOINC 6.10.17. Here is an example workunit:
https://secure.worldcommunitygrid.org/ms/devi...s.do?workunitId=195756488 I left this running last night -- it showed 15% done with just under 6 hours of CPU time (I know the %done calculation is never right). Looked at the results this morning, and it reported less than 3 hours of CPU time, even though it had nearly 6 under its belt before I left it. |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
OK, here is an actual comparison between what I saw for a completed workunit just after completion on my computer, and what I see after upload in "Results Status"
----------------------------------------Here's the link to the workunit: https://secure.worldcommunitygrid.org/ms/devi...s.do?workunitId=195270890 [Edit 1 times, last edit by Former Member at Oct 11, 2010 1:53:27 AM] |
||
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
gretchch,
----------------------------------------We cannot see information on webpages to results you link to. Copy paste will work, but we believe you. See post by armstrdj. There's full awareness on lost CPU time, though the result computes properly. There's 2 options. 1. Continue crunching CEP2 on the affected device(s) 2. Disable CEP2 in your device profile until the word comes back that the time recording is fixed. I've no answer why this flared up for some but not all. Suspicions from my side are a Linux patch aggravated the situation. And to reconfirm the reconfirmed, my Lucid Lynx 10.04.1 quad does not show these big losses, at all, not on a single result. In that I wonder from another conversation, anyone having the Linux firewall on? If so is port 31416 over localhost (127.0.0.1) open? I've tried in UFW to set a general exception for BOINC, but been unsuccessful, so switch the FW off again. I'm behind a hardware based FW anyhow.
WCG Global & Research > Make Proposal Help: Start Here!
Please help to make the Forums an enjoyable experience for All! |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I will continue crunching, just wanted to provide a concrete example.
FYI -- I have an AMD Phenom 2.4 GHz quad-core in the machine I gave results for above. I have another Linux box also running 32-bit Ubuntu 10.04 that has Intel processors, which I suspect has the same problem, since all it's workunits are now coming in under 3 hours while its wingman's workunits are all over 5 hours. This other box is an old/slow Thinkpad, so it should definitely not be crunching faster than most other machines. My AMD quad core does not have firewall enabled, the Intel box does have firewall enabled. Both boxes are 64-bit capable, even though they have 32-but Ubuntu installed, so they both have the PAE kernel. About the only other thing they have in common (besides the obvious Ubuntu distro) is ClamAV. |
||
|
mwgiii
Advanced Cruncher United States Joined: Aug 17, 2006 Post Count: 131 Status: Offline Project Badges: |
I just noticed that my CPU times are now mostly under 3 hours while run times are still in the 9 hour range. This is down from around 7 hours CPU time per WU.
----------------------------------------Looking back through my completed units, my CPU time took a nosedive on 9/27/10. |
||
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
So it seems it started more pronounced for different members on different dates. Maybe it can be connected back to a specific batch. Depending on cache sizes it is reasonable to think the date will vary, but what else happened to the OS/Host/client when this started.
----------------------------------------Saw now twice CalmAV mentioned (in Beta and in production), which seemingly is an active type of AV for Linux. That again brings up the question of excluding scanning of the BOINC data_dir and science apps. Notably I just killed the Linux UFW firewall. It was impacting my internet download speed of files of anything over 1MB, even LAN device-device synchronizations went crawling. Can't though say it's the root cause, UFW just being the symptom testing some other weakness.
WCG Global & Research > Make Proposal Help: Start Here!
Please help to make the Forums an enjoyable experience for All! |
||
|
mwgiii
Advanced Cruncher United States Joined: Aug 17, 2006 Post Count: 131 Status: Offline Project Badges: |
I'm running Ubuntu using VMWare in Windows Vista on my quad. So I am expecting a difference between CPU time and run time since 2 cores are being shared. I am only running CEP, WUProp, and FreeHAL in Linux.
----------------------------------------It very well could be a Ubuntu patch which was installed on my machine on 9/27 but was released a couple of days earlier because I don't run VMWare 24/7 because VMWare absolutely kills my GPU runtimes. I don't have ClamAV running. [Edit 1 times, last edit by mwgiii at Oct 12, 2010 5:08:53 PM] |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I have a 2 hour cache, in the hopes of pulling in the occasional Beta WU. My problems seems to have started on 30 September. The CEP2 workunit just before my first small one was an Error:
Result Name: E200366_ 613_ A.24.C19H12N2OS2.271.4.set1d06_ 2-- <core_client_version>6.10.17</core_client_version> <![CDATA[ <message> WU download error: couldn't get input files: <file_xfer_error> <file_name>E200366_A.24.C20H13NOS2.114.2.zip</file_name> <error_code>-224</error_code> <error_message>file not found</error_message> </file_xfer_error> </message> ]]> After that, All of my workunits show up as under 3 hours, starting with E200397_ 705_ A.25.C18H11N5OS.44.4.set1d06_ 0-- |
||
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Doubt it helps, but a project reset will force a clean download of all permanent science application files. Not even an uninstall, reinstall does that... it simply continues with the same files that were already in the data_dir.
----------------------------------------
WCG Global & Research > Make Proposal Help: Start Here!
Please help to make the Forums an enjoyable experience for All! |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I just checked my Synaptic package manager history, and there were a LOT of updates that went on to my box on 29 September:
Commit Log for Wed Sep 29 22:13:00 2010 Upgraded the following packages: avahi-autoipd (0.6.25-1ubuntu6) to 0.6.25-1ubuntu6.1 avahi-daemon (0.6.25-1ubuntu6) to 0.6.25-1ubuntu6.1 avahi-utils (0.6.25-1ubuntu6) to 0.6.25-1ubuntu6.1 cheese (2.30.1-0ubuntu1) to 2.30.1-0ubuntu2 cheese-common (2.30.1-0ubuntu1) to 2.30.1-0ubuntu2 gnome-terminal (2.29.6-0ubuntu5) to 2.30.2-0ubuntu1 gnome-terminal-data (2.29.6-0ubuntu5) to 2.30.2-0ubuntu1 google-chrome-stable (6.0.472.62-r59676) to 6.0.472.63-r59945 libavahi-client3 (0.6.25-1ubuntu6) to 0.6.25-1ubuntu6.1 libavahi-common-data (0.6.25-1ubuntu6) to 0.6.25-1ubuntu6.1 libavahi-common3 (0.6.25-1ubuntu6) to 0.6.25-1ubuntu6.1 libavahi-core6 (0.6.25-1ubuntu6) to 0.6.25-1ubuntu6.1 libavahi-glib1 (0.6.25-1ubuntu6) to 0.6.25-1ubuntu6.1 libavahi-gobject0 (0.6.25-1ubuntu6) to 0.6.25-1ubuntu6.1 libavahi-ui0 (0.6.25-1ubuntu6) to 0.6.25-1ubuntu6.1 libcheese-gtk18 (2.30.1-0ubuntu1) to 2.30.1-0ubuntu2 libgdiplus (2.4.2-1build1) to 2.4.2-1ubuntu0.10.04.1 libmikmod2 (3.1.11-a-6.1) to 3.1.11-a-6.1ubuntu0.1 libphonon4 (4:4.6.2-0ubuntu5) to 4:4.6.2-0ubuntu5.1 libqt4-assistant (4:4.6.2-0ubuntu5) to 4:4.6.2-0ubuntu5.1 libqt4-dbus (4:4.6.2-0ubuntu5) to 4:4.6.2-0ubuntu5.1 libqt4-designer (4:4.6.2-0ubuntu5) to 4:4.6.2-0ubuntu5.1 libqt4-help (4:4.6.2-0ubuntu5) to 4:4.6.2-0ubuntu5.1 libqt4-network (4:4.6.2-0ubuntu5) to 4:4.6.2-0ubuntu5.1 libqt4-opengl (4:4.6.2-0ubuntu5) to 4:4.6.2-0ubuntu5.1 libqt4-qt3support (4:4.6.2-0ubuntu5) to 4:4.6.2-0ubuntu5.1 libqt4-script (4:4.6.2-0ubuntu5) to 4:4.6.2-0ubuntu5.1 libqt4-scripttools (4:4.6.2-0ubuntu5) to 4:4.6.2-0ubuntu5.1 libqt4-sql (4:4.6.2-0ubuntu5) to 4:4.6.2-0ubuntu5.1 libqt4-sql-mysql (4:4.6.2-0ubuntu5) to 4:4.6.2-0ubuntu5.1 libqt4-sql-sqlite (4:4.6.2-0ubuntu5) to 4:4.6.2-0ubuntu5.1 libqt4-svg (4:4.6.2-0ubuntu5) to 4:4.6.2-0ubuntu5.1 libqt4-test (4:4.6.2-0ubuntu5) to 4:4.6.2-0ubuntu5.1 libqt4-webkit (4:4.6.2-0ubuntu5) to 4:4.6.2-0ubuntu5.1 libqt4-xml (4:4.6.2-0ubuntu5) to 4:4.6.2-0ubuntu5.1 libqt4-xmlpatterns (4:4.6.2-0ubuntu5) to 4:4.6.2-0ubuntu5.1 libqtcore4 (4:4.6.2-0ubuntu5) to 4:4.6.2-0ubuntu5.1 libqtgui4 (4:4.6.2-0ubuntu5) to 4:4.6.2-0ubuntu5.1 linux-generic-pae (2.6.32.24.25) to 2.6.32.25.27 linux-headers-generic-pae (2.6.32.24.25) to 2.6.32.25.27 linux-image-generic-pae (2.6.32.24.25) to 2.6.32.25.27 linux-libc-dev (2.6.32-24.43) to 2.6.32-25.44 phonon (4:4.6.2-0ubuntu5) to 4:4.6.2-0ubuntu5.1 python-avahi (0.6.25-1ubuntu6) to 0.6.25-1ubuntu6.1 python-mako (0.2.5-2ubuntu1) to 0.2.5-2ubuntu1.3 python-software-properties (0.75.10) to 0.75.10.1 software-properties-gtk (0.75.10) to 0.75.10.1 software-properties-kde (0.75.10) to 0.75.10.1 Installed the following packages: linux-headers-2.6.32-25 (2.6.32-25.44) linux-headers-2.6.32-25-generic-pae (2.6.32-25.44) linux-image-2.6.32-25-generic-pae (2.6.32-25.44) |
||
|
|